We have to make a lot of decisions in our lives, but sometimes when given two choices, going with both can be the right answer. For example, researchers working on neurodevelopment figured out two different phenomena that regulate non-CG methylation in the postnatal brain, but how are they connected? Now, Harrison Gabel’s lab (Washington University School of Medicine in St. Louis) reports that two phenomena—gene topology and gene expression—are tied together through NSD1, which directs non-CG methylation to the right places.
Shortly after birth, neurons are still making up their minds, undergoing further epigenetic changes. DNMT3A is putting down non-CG methylation—mostly at CA (mCA)—and MeCP2 reads these marks to regulate enhancers and genes. Disrupting either DNMT3A or MeCP2 can lead to neurodevelopmental disorders (NDDs), such as Tatton-Brown-Rahman syndrome (TBRS). Although NSD1 disruption leads to another NDD called Sotos syndrome, no one really knew how NSD1 fits into the mCA picture.
Here are the two phenomena that researchers knew about:
- A lot of transcription means less DNMT3A binding and less mCA
- mCA levels are different throughout the genome, depending on regions of topologically associating domains (TADs), with high-set-point TADs having a lot more DMNT3A binding and more mCA
So, Gabel’s team decided to see how these observations could be unified, looking at postnatal mouse brain development. They used many methods, including Hi-C, RNA-seq, whole-genome bisulfite sequencing, ChIP-seq, and found:
- TAD set points and transcription levels were independent, but their effects on mCA levels added up to influence the level of MeCP2 regulation
- Just as in other parts of the body, H3K36 methylation interacted with DNMT3A in the postnatal brain, with DNMT3A binding H3K36me2 and correlating with lower expression; highly expressed genes had more H3K36me3 and less DNMT3A
NSD1 is mutated in Sotos syndrome and has similar symptoms to TBRS, in which DNMT3A is mutated, so the team chose to look in more depth at NSD1 and observed:
- In vitro NSD1 disruption altered H3K36me2, DNMT3A and mCA signals, which appeared to be affected by TADs; so NSD1 is needed for H3K36me2 patterns, which affects mCA set points and levels
- An NSD1-conditional knock-out mouse showed similar results to the in vitro experiment and revealed increased transcription as H3K36me2, DNMT3A, and mCA levels declined
Taken together, the data indicates that gene topology causes NSD1 to deposit H3K36me2 throughout the genome and represses transcription. But some H3K36me2s selectively get converted to H3K36me3, which doesn’t bind DNMT3A well, alleviating repression and increasing expression of some genes and elements. This conversion decision means less DNMT3A binding and less mCA where transcription is high.
Once you’ve made your decision to dig into the details, read more at Molecular Cell, May 2023.